Upon biosynthesis at the ribosomal complex, proteins undergo a series of editing steps through which proper folding and/or proper modification (e.g. by attachment of carbohydrate chains, lipid chains, or by phosphorylation) are ascertained. The editing steps occur in a stepwise manner on the surface, or within, highly specialized vesicular structures that are part of a highly dynamic network of vesicles within the cytoplasm, termed endoplasmic reticulum. In order to guide proteins through the editing processes, unique elements within the protein structure have evolved that act as sorting signals, by facilitating insertion of a given protein into membranes (a well-recognized function of attached lipid or carbohydrate residues), by defining interaction with so-called chaperones (transport molecules that move a given protein into or from a subcellular environment), or by a combination of such mechanisms. In essence, specific amino acid sequences acting as sorting signals match a given protein to its final destination where it is to exert its function, e.g. the plasma membrane for a surface receptor, the nucleus for a transcription factor, the mitochondrium for an enzyme participating in the respiratory chain. Many principles of sorting have been elucidated by the seminal studies of Blobel who was awarded the Nobel Prize for Medicine and Physiology in 2001. A large number of research groups have assembled the current knowledge about sorting signals for individual organelles (see Trends in Biochem Sci 16, 478-481, 1991 for an overview over nuclear translocation sequences; see Hartl and Neupert, Science 247, 930-938, 1990; Pfanner et al., Annu Rev. Cell Dev. Biol. 13, 25-51, 1997 for review of mitochondrial translocation sequences). The details of sorting are highly complex and differ with cellular environments, and most aspects of sorting, especially how multiple sub-elements of a sorting signal function, or how multiple signals for different destinations within a protein structure are assigned their order and priority of processing, are currently not understood.
In particular, the details of mitochondrial import of proteins are still unclear. The inner space of mitochondria is surrounded by two distinct membranes. Within each membrane, a designated protein complex acts as gatekeeper for macromolecules crossing the membrane. There is no “consensus sequence” of mitochondrial import signals that could act as reference for studies on structure-function-relationship. Most mitochondrial import signals are of higher complexity and are encoded by relatively long peptide sequences. After import, the signal sequence of naturally occurring proteins is cleaved off by the translocase. The recent crystal structure analysis of the enzyme importing macromolecules into the inner space of mitochondria has revealed some structural insight about the structural requirements on import sequences. Repeats of short helical structure are facilitating the translocation step proper, whereas the protein region encoding the import signal has to undergo a conformational change into a more unfolded structure to permit proteolytic cleavage off the imported protein. The sum of the individual processes of guided intracellular movements of a given macromolecule from one environment to another is often collectively referred to as “trafficking”.
The invention contemplates to reduce drug toxicity and to increase drug efficacy by imparting the trafficking pathways of endogenous macromolecules upon therapeutic and imaging drugs designed to exert their function within a specific subcellular environment.
For the sake of clarity, definitions are given as follows (see also FIG. 2 and the legend to FIG. 2 in co-pending application U.S. Ser. No. 09/428,675 filed Oct. 27, 1999): A “presentation molecule” encompasses an extracellular targeting moiety, connected to an intracellular targeting moiety, termed “routing moiety”, that is carrying a “bio-active molecule”, or BAM.
The extracellular targeting moiety acts as a ligand (preferably a non-agonist ligand) for a transmembrane receptor that participates in internalization (a process wherein a membrane protein leaves the membrane environment, releases bound ligands to the intracellular space, and eventually returns to the membrane environment). The “routing moiety” participates in intracellular trafficking and delivers the drug molecule to a pre-determined subcellular environment. The “bioactive molecule” is the pharmacologically most active element in the composition and exerts the therapeutic effect, and/or the imaging function. In the case of a dual modality PET(SPECT)/BNCT drug, the routing moiety and the bio-active molecule are fully integrated, generating a unique drug delivery molecule (DDM). “Subcellular environment” is the collective term for all cellular organelles (such as nucleus, mitochondria, etc.), vesicular networks (such as the endoplasmatic reticulum with the Golgi apparatus), dynamically changing vesicular complexes (such as lysosomes and vesicles encountered in trafficking), and the cytoplasmic space between organelles and vesicular elements.
The use of 10B which when subjected to epithermal neutrons decays to lithium-7 and an alpha particle is well known and has been suggested previously for destruction of cancerous cells, (see Inhibition of Human Pancreatic Cancer Growth in Nude Mice by Boron Neutron Capture Therapy, Hyyanagie, et al., British Journal of Cancer, 75 (5), 660-665 (1997). It is frequently referred to by the acronym BNCT. The potential efficacy of BNCT for malignant glioma is discussed in Boron Neutron Capture Therapy: Implications of Neutron Beam and Boron Compound Characteristics, F. J. Wheeler, et al., Med. Phys. 26 (7), 1237-1244 (July, 1999). Two articles by Y. Imahori, et al. discuss PET based BNCT using boron compound labeled with 18F, see Clinical Cancer Research, 4, 1825-1841 (August, 1998). U.S. patents relating to the use of BNCT include U.S. Pat. Nos. 6,074,625, 6,248,305 and 5,846,741.
In essence, BNCT employs a stable non-radioactive isotope such as boron-10 (10B) which upon capturing a thermal neutron causes a fission reaction. In the 10B fission, the resulting alpha and lithium particles have high energy, LET and RBE and travel less than 10 microns in tissue. As a result, selective tumor-cell killing is provided. An external beam of epithermal neutrons, with an energy maximum between 100 and 1000 eV, is employed to treat the 10B-loaded tissues by focusing upon a specific region in the body where the treatment is desired, and thus avoiding activation of the boron isotope that may have been retained in other parts of the body which may also contain the receptors being targeted. Details of reactor design and neutron beam validation are given in Wheeler et al., Med. Phys. 26, 1237-1244, 1999. Lee et al. in Med. Phys. 27, 192-202, 2000, describe a modified accelerator that is portable, more economical, and could make BNCT available for widespread hospital use. Alburger et al. in Med. Phys. 25, 1735-1738, 1998, have developed highly sensitive phantoms for thermal neutron depth profiling that can be used for validation of neutron beams obtained from accelerator-based BNCT facilities. In absolute numbers, a 10B content of 5-30 ppm in tumor cells is necessary to cross the threshold for effective BNCT, corresponding to a number of about 109 10B atoms distributed uniformly throughout a tumor cell (Fairchild and Bond, Int. J. Radiat. Oncol. Biol. Phys. 11, 831-840, 1985). The desirable therapeutic range between 5 and 30 microgram/g tumor tissue (Coderre and Morris, Radiat. Res. 151, 1-18, 1999). Subsequent calculation of tissue dosimetry has been conventionally accomplished by Monte Carlo simulations of cell destruction using estimated intracellular and extracellular concentrations of the radiopharmaceutical (Kobayashi and Kanda, Radiation Research 91, 77-94, 1982). The accuracy of such estimations can be substantially improved by utilization of a boron compound that is accumulating, and detectable by a second imaging modality, in a pre-determined cellular microenvironment, preferably the nucleus where neutron capture will exert the best therapeutic effect.
Monitoring and imaging of gene expression upon gene therapy has become an important procedure for validation of experimental gene therapy regimens. Most gene repair efforts have to be targeted to a specific organ, such as the liver or intestinal epithelia. Both temporal and spatial parameters of gene expression are to be assessed in living organisms in real-time mode to evaluate the effectiveness of the chosen gene delivery procedure. A variety of imaging techniques has been tested. Weissleder et al. (Nature Medicine 6, 351-355, 2000) describe an experimental procedure wherein a tumor made to overexpress an engineered version of the transferrin receptor could be visualized in vivo by transferrin-linked paramagnetic particles detected by MRI. However, the increase of paramagnetic particle uptake was only 2.5-fold compared to tumors not expressing the engineered receptor gene, and it was concluded by Weissleder et al. that temporal resolution was limited, and probe detection sensitivity was several orders of magnitude lower than in alternative imaging modalities, such as optical imaging and imaging of nuclear isotopes.
As alternative imaging modality, PET scanning has been tested successfully. Certain isotopes emit positively charged particles of a mass close to zero (positrons) that otherwise have the wave properties of negatively charged electrons. If a positron and an electron collide, each particle undergoes conversion into a gamma ray of 511 keV energy; since both gamma rays are emitted into opposite directions at an angle of 180 degrees, it is feasible to scan such conversion events as coinciding gamma rays in paired detectors using e.g. lutetium oxyorthosilicate as scintillation detection material, while eliminating those gamma rays that do not coincide. Details of current PET instrumentation are described in Fahey, Radiol. Clinics North Am 39, 919-929, 2001.
Because of high cost in running dedicated PET systems, a less expensive imaging modality (single photon emission computed tomography, SPECT, also abbreviated SPET, for single photon emission tomography) has recently gained popularity whereby a gamma ray in the energy range of 30 to 300 keV energy is emitted and detected by a modified dual-head, or multiple head, gamma camera system. SPECT imaging can be performed with isotopes of longer half-life than those used in PET, such as 111In or 99mTc, that are well-characterized in nuclear medicine and can be shipped from dedicated radiochemistry facilities. SPECT imaging of leukocytes labeled with 111In or 99mTc has been validated as “gold standard” in detection of occult infectious and inflammatory sites (Renken et al, Eur. J. Nucl. Med 28, 241-252, 2001). While the use of conventional gamma camera technology is somewhat of an advantage for image acquisition, conventional SPECT does not employ electronic collimation of incoming gamma rays (Shao et al, Phys. Med. Biol. 42, 1965-1970, 1997), and thus is estimated to have lower sensitivity than PET by at least one order of magnitude. Multiple head detection systems and advanced image construction software are critical to achieve optimal imaging. SPECT imaging has been found to be a safe and cost-effective method with advantages over CT and other imaging methods in diagnosis and management of lung cancer patients (Goldsmith and Kostakoglu, Radiol. Clinics North Am. 38, 511-524, 2000). Details of current SPECT instrumentation, especially novel useful SPECT/PET hybrid detection systems, are described in Fahey, Radiol. Clinics North Am 39, 919-929, 2001.
Increasing the concentration and limiting the source of SPECT in a defined microenvironment within a target cell by pharmacological means would be a significant improvement, both for image resolution and for radiation planning by other modalities such as BNCT. It is a particular disadvantage and source of error if equal distribution of a radiation source across all cell compartments has to be assumed, rather than measuring it directly in the microenvironment that is important for the intended radiation therapy.
It has been proposed that SPECT and PET are useful to perform imaging in mouse models, with a resolution of about 1 to 2 mm and a signal collection time in the range of minutes (reviewed by Weissleder, Nature Reviews in Cancer 2, 2002); a mouse PET system to permit 1 mm resolution has been reported as in development by Hershman et at (J. Neuroscience Res. 59, 699-705, 2000). Alternative imaging modalities include the use of fluorescent reagents (Honigman et al, Mol. Therap. 4, 239-249, 2001) and Yang et al. Proc Natl Acad Sci 98, 2616-2621, 2001); the clinical use is still limited, because non-invasive detection is limited to pathological processes close to the surface, with a maximal depth of 10 cm in fluorescence-mediated tomography (Weissleder 2002).
Ray et al. have published a synopsis and a review about the application of PET to monitor gene therapy (see Table 1 in their publication in Sem. Nucl. Med. 31, 312-320, 2001). The best characterized model is by way of delivering a gene from Herpes Simplex virus encoding a thymidine kinase that will accept as substrate synthetic derivatives of uracil (e.g. 2′18F-2′-deoxy-1-beta-D-arabinofuranosyl-5125I-uracil) and guanosine (e.g. 18F-ganciclovir or 18F-penciclovir), whereas the naturally occurring thymidine kinase does not. The level of gene expression to be tested is proportional to the amount of phosphorylated uracil derivative or guanosine derivative that is retained intracellularly upon phosphorylation and can be detected by PET scan. Similarly, the gene encoding somatostatin receptor type 2 has been delivered to cells to permit imaging by PET (Rogers et al., Q J Nucl Med 44, 208-223, 2000). This approach is still bound to the limitations of having an extracellular ligand contact a target of unknown and possibly low surface density, and cannot discriminate between cells naturally expressing the somatostatin receptor and those targeted successfully by gene therapy. Neuroendocrine tumors and gastrointestinal tissues express somatostatin type 2 receptor naturally; a tangible benefit for imaging by raising the level of somatostatin receptor further may apply to only a small subset of tumors. Furthermore, gene therapy to the liver may not be reportable at all. The Herpes virus thymidine kinase approach is far superior, because a detectable artificial intracellular substrate is enzymatically enriched only in cells that express the transgene at sufficiently high density. Still, it would be a significant improvement if the administration and overexpression of a viral enzyme interfering with energy metabolism and DNA synthesis could be avoided.
Mitochondria play a major role in the metabolism of eukaryotic cells and control pathological processes in disease and aging. For example, impaired mitochondrial function has a direct impact on ATP synthesis, regulation of intracellular Ca++ homeostasis, generation of free radicals, and execution of programmed cell death pathways. It is therefore of great interest to target pharmaceutical compositions to mitochondria, for the purpose of monitoring physiological are pathological processes, or to deliver therapeutic drugs (see Murphy, Trend in Biotechnol. 15, 326-330, 1997). Specifically, tumor cells are characterized by a higher cell membrane potential and also a higher mitochondrial membrane potential (Chen, Annu Rev. Cell Biol 4, 155-181, 1988), permitting compounds like lipophilic cations to accumulate with a certain selectively in tumor cell mitochondria. The disruptive effect of lipophilic cations may be exerted by increase of the proton permeability of the inner membrane and inhibition of respiration (Azzone et al, Curr. Topics Bioenerg. 13, 1-77, 1984), or by more specific inhibitory effects e.g. on mitochondrial transcription. Lipophilic cations have been used as carriers to deliver the cytotoxic compound cisplatin to tumor cells (Steliou, U.S. Pat. No. 6,316,652). Furthermore, cation conjugates have been designed that combine to form a toxic product once inside the cancer cell mitochondrium (Rideout, Cancer Invest 12, 189-202, 1994).
Zhang and Haugland (U.S. Pat. No. 5,686,261) have disclosed fluorescent substituted 3′-6′-diaminoxanthenes which selectively localize in mitochondria and are retained after fixation, permeabilization, and cell death. No pharmaceutical applications are contemplated. The use of full-length bovine heart mitochondrial protein DNA sequence is suggested for creating a fusion gene with a yeast gene of interest, which upon expression will be imported into mitochondria. Herrnstadt et al. (U.S. Pat. No. 6,171,859) have disclosed methods and compositions to destroy mitochondria with defective cytochrome c oxidase in patients with Alzheimer's disease by way of a toxin conjugated to targeting molecule which is a lipophilic cation. The toxin may be a small-molecule agent or an antisense oligonucleotide. Alternatively, an imaging ligand (e.g. radioisotope suitable for PET or SPECT) may be coupled to the targeting molecule for in vivo imaging of defective mitochondria which will accumulate the imaging drug through increased membrane potential and increased levels of negatively charged phospholipids.
Because tumor cells inside of a tumor are often without significant oxygen supply, the mitochondrial membrane potential is not defined by respiration, and thus may not be optimal for uptake of lipophilic cations. Milder forms of oxygen deprivation in the periphery of tumors may elicit a stress response leading to the observed hyperpolarization of cancer cell mitochondrial and plasma membranes. Efficacious lipophilic cation uptake is thus limited to a small portion of the tumor cells. Another limitation is the access of lipophilic cations to mitochondria in normal cells which represent a pool vastly in excess over cancer cell mitochondria. Alternative modes of mitochondrial targeting are very desirable.
The use of mitochondrial import sequences instead of lipophilic cations might improve delivery of a drug to a larger population of mitochondria in a tumor, but would at the same time target mitochondria in non-tumor cells as well, thus losing a critical element of selectivity. Again, alternative modes of mitochondrial targeting are very desirable.
Somatostatin receptor ligands have been used to image cells by visualizing the level of somatostatin receptor present on the surface of a cell in a disease condition. Specific radiolabeled somatostatin analogs are disclosed in EP 607103 to visualize somatostatin receptor in primary tumors, tumor metastases, cells affected by inflammatory and autoimmune disorders, tuberculosis, and cells in organ rejection after transplantation. The use of somatostatin analogs for imaging somatostatin receptor positive cells and tissues, in particular tumors, metastases, inflammatory disorders and autoimmune disorders, is further disclosed in WO 97/01579. U.S. Pat. No. 5,976,496 discloses the use of somatostatin analogs to image atherosclerotic plaque, in particular non-critically stenotic plaque and unstable atherosclerotic plaque.
Neither of the references suggests that somatostatin analogs may be used to image angiogenesis by way of a tripartite molecule comprising a non-agonist targeting moiety, an intracellular routing moiety, and a radionucleid linked to the routing moiety. In the current invention, the somatostatin analog is not derivatized to carry a radionucleid. Neither of the references suggests using the somatostatin receptor as port of entry, translocating the imaging drug to a pre-determined subcellular microenvironment by way of sorting signals, and attaching a radiolabel to a molecular target other than a somatostatin receptor.